---
id: 20260812-T0-06
title: "动态联盟与通信定价：多智能体AI系统协作效率新框架"
title_en: "Dynamic Coalition Formation and Communication Pricing Framework for Agentic AI Systems"
url: https://ai.daily.yangsir.net/daily/20260812-T0-06
issue_date: 2026-08-12
publish_date: 2026-08-11T04:00:00.000Z
category: research
source_name: "arXiv cs.AI"
source_url: https://arxiv.org/abs/2608.07532
---

# 动态联盟与通信定价：多智能体AI系统协作效率新框架

arXiv新论文提出一种针对技能异构的多智能体AI系统的动态联盟形成与通信定价机制。现有架构要么固定通信拓扑，要么全广播，导致token成本、延迟和冗余过高。新方法让智能体根据任务动态组队，并对通信资源合理定价，在保持系统性能的同时显著降低通信开销。

## English Version

**Dynamic Coalition Formation and Communication Pricing Framework for Agentic AI Systems**

New paper proposes dynamic coalition formation and communication pricing for multi-agent AI systems. Fixed communication or full broadcast wastes tokens and latency. The framework lets agents form teams dynamically and price communication, cutting overhead while maintaining performance.

---

**来源**：[arXiv cs.AI](https://arxiv.org/abs/2608.07532)

**详情页**：https://ai.daily.yangsir.net/daily/20260812-T0-06

---

*智语观潮 · Daily — https://ai.daily.yangsir.net/llms.txt*